The objective of this research was to detect protein content of single rice kernel using near-infrared technique. The near-infrared multi-spectral imaging system with detection wavelength at 890, 910, 920, 990, and 1014 nm was adopted for measuring protein content of single rice kernel. The rice samples were collected from ten different cultivars to determine the protein content. The multiple linear regression (MLR) analysis was adopted for predicting rice protein content by near-infrared multi-spectral imaging. The coefficients of determination (r2) for the calibration and validation sets based on 890, 910, 920, 990, and 1014 nm were 0.821 and 0.464 respectively. The result shows that it is feasible to use near-infrared technique to determine protein content of single rice kernel in the rice milling processing.